Solenoid Electromagnet with strong core, low residual magnetism

  • Thread starter rsr_life
  • Start date
  • #1
Hello, this is my first post here.

I'm building an electromagnet as part of a larger experiment in my lab. My magnet needs to move a magnetic thread made with a paramagnetic substance. This string just barely responds to strong magnetic fields (tested it with standard speaker magnets). The electromagnet I'm building thus needs to generate a strong enough magnetic field to move this piece, even just so slightly. I can't afford to use a ferrite core because of hysteresis effects - I need to rapidly switch between multiple electromagnets and cant afford to have hysteresis causing a delay. Any suggestions on how I can do this? Suggestions on any core that I can use to amplify the field that has a really low coercivity? or any configuration (i'm currently trying a regular solenoid, maybe horseshoe solenoid?, dont know) or even any other method that I can employ to generate the field.

To summarize, I need to switch between multiple electromagnets to move a magnetic material (that responds poorly, hence a strong field is required) and need a solution on magnifying the field intensity that doesn't involve expensive cores (mumetal costs a lot, i think).

Big question, I'm sure. I'm using a Gauge 20 wire and have a standard power supply for power (about 3 amps max).

Tried higher gauge wires - more turns, easier to wind - but they heat up too much and the field isn't impressive anyway.

Appreciate any help in advance,


Answers and Replies

  • #2
Well there are lots of materials (different ferrites, et. al.)
that are used in magnetics up into the MHz and GHz ranges
of frequencies, so it is by no means impossible to have them
respond quickly enough to follow fast changing signals
despite their hysteresis.

The bigger problem with fast changing fields is that the
materials that respond more quickly (ferrites, powdered
iron, et. al.) tend also to have much lower permeabilities.
Ferrites, specifically, also tend to have relatively low
saturation magnetic field strengths also. So though they
can respond quickly, there's a fairly low limit as to the
magnetic field strength they can generate.

If you're looking for a magnetic field from several thousand
gauss to perhaps in the near-1-tesla range, AND you need
to change the field quickly, it's often preferable to just
use good old air core electromagnets with high
Amp-turn values. That way the hysteresis will be
nearly zero, the inductance can be made manageably low,
and there's no non-linear limit to the attainable field
strength. Also the wire heating will be limited by the
fact that you're seemingly operating in pulse mode
where any given coil / electromagnet will be turned on
then off quickly. In the place of a core use a water
cooling setup to jacket and wind the coils around if
further cooling is needed.

Take a few big deep cycle marine / golf cart batteries,
some relatively heavy gauge magnet wire (12g?), wind a
few turns so that the resistance and inductance limits the
current to something thermally managable (12G
melts at ~ 235A in open air convection, so assume
you're not going to sustain more than around 100A for
long even with flowing cold water cooling), and see
how that works for you.

Use an appropriately rated IGBT for switching the currents,
and use much heavier gague wire for the connections
from the battery terminals to the coils -- like 4G or 0G
or whatever .. moderately heavy automotive jumper
cables are about right.

You'd end up with something like 400 Amp-turns or more
over a ring that could be an inch in diameter or less;
that'd probably get the attention of your paramagnetic

For a test just get a car battery and a pair of jumper
cables and wind a little coil out of 12G solid wire from
some household ROMEX or something and use heat-shrink
tubing or electrical tape to insulate it and manually
use the jumper cables to generate a swiping contact
of less than a second and see if it reacts well.

I don't know how long your pulses are to be, or what your
dimensional field requirements are, et. al. or I'd venture
to give you more specific coil winding data.

Also just because it's paramagnetic doesn't mean that
that's the best way to mechanically couple to it to cause
motion -- consider possible electrostatic effects depending
on its dielectric properties, as well as things like using
air bursts or mechanical vibrations et. al.
  • #3
Thanks for that XEZ,

I understand what you're saying that I do. Unfortunately, the experimental settings require that the current injected be of a fairly low value, given that I'll be using a DAC board set to current output. Haven't specified which DAC board i'll be using, but i believe the current output is still in the 1 or 2 amps range for the DACs we have around here. And thats one of the problems, the current output has to be that low for applications like these and that compounds issues elsewhere. I'm using the 20G wire currently.

The pulse length depend on how quickly i can get the thread to respond to my field. If it responds quickly, and the hysteresis is low, then that would mean the pulse can be short. Would a field be generated instantly for a 20G wire with low amp? Can I use some kind of Soft magnetic material? Soft ferrite maybe?

  • #4
Well the rate at which the field establishes when you're
using a magnetic core material and electromagnet will
depend on a couple of things -- a) the core material's
hysteresis curve and physical volume will dictate how
much energy you have to apply to switch it in a given time.
b) the inductance of your coil will cause the current
to rise at a rate depending on your voltage.

1 Henry = 1 Volt * Second / Amp, so
10V for 1 second would cause the current to ramp by 1A.
1 Henry also equals one Weber / Amp.

Magnetic flux density:
1 Tesla = the strength of a very strong Neodymium
rare earth magnet, probably many times stronger than
the speaker magnet you're likely to have used.
At 1 Tesla fields you're hard pressed to be able to pull
a magnet off of an iron surface if it's even possible;
it's uncontrollably strongly attracted within a few
millimeters of an iron surface, and it'll accellerate so
quickly together it'll either shatter or throw sparks.

1 Tesla = 1 Weber / square meter

B [in Tesla units] = u_r * u_0 * H

Magnetic field intensity:
H = Amps * Turns / meter

u_r = relative permeability of your core as a ratio
compared to free space.

u_0 = permeability of free space = 4 * Pi * 10^-7

So since your speaker magnet's insufficient, I assume
you're talking about flux densities of around
0.25 to 1 Tesla.

Those kinds of fields are just about impossible to generate
in a compactly sized soft electromagnet operating at
3 Ampere current levels; you'd need a lot of turns and
a moderately sizable core volume of quite high
permeability and saturation field strength. It's very
possible, but when you start using dozens to hundreds
of turns and high permeability cores your inductance
and hysteresis energy becomes such that rapid pulsing
(e.g. several kilohertz) becomes non-trivial in power
requirements and hence voltage and current levels.

What are the rough physical dimensions you're able to use
for a single pole-piece of the magnet, and then how
much bigger and in what proximity can the rest of the
electromagnet be? How close will the pole piece get
to the sample, and what's the needed area of pole piece
to become comparable to or larger than the area of the

Anyway I'll take a look at a few of my references to see
if I can find materials that have saturation flux
densities in the 0.25 to 1.0 Tesla range, and
relative permeabilities in the 10^3...10^4...10^5 range.

Certainly many of the classic ferromagnetic materials
meet the saturation field and permeability requirements,
but they're not generally appropriate for <= 100uS
switching unless you apply a LOT of electrical power.

Could you just use a Nd rare earth magnet and use
the DAC to control a voice coil (e.g. speaker, galvanometer,
or alternatively a piezo stack) to physically *MOVE* the
small powerful rare earth magnet to modulate its
field intensity at the thread's location at reasonably
high frequencies? A little creativity with a stereo
amplifier and speaker could get your mechanical
magnet motion thing going.
  • #5
This string just barely responds to strong magnetic fields (tested it with standard speaker magnets). The electromagnet I'm building thus needs to generate a strong enough magnetic field to move this piece, even just so slightly.
Just a thought, but why use a magnetic field if the string just barely responds to a magnetic field. Why not move it some other way? Even if it has to be non-contact, just blast it with compressed air or something?
  • #6
Just a thought, but why use a magnetic field if the string just barely responds to a magnetic field. Why not move it some other way? Even if it has to be non-contact, just blast it with compressed air or something?
Hey, thanks a bunch for that info XEZ,

The sample that I mentioned has a diameter of around 40 microns but its about 30 cms in length - like a long thread. Its pretty light and I believe it's used for microfluidics applications - so thats the scale I'm talking about. It weighs less than a gram, and just one tip needs to be moved maybe 1 cm or so. The other electromagnets need to be close by (same distance) for the switching to move the sample this distance.

The idea you suggested is creative. The ones that I've tried involve reconfiguring the whole setup physically to then induce some tiny motion magnetically. Is a Nd magnet available off-the-shelf? I understand that the requirements would demand stronger, more powerful magnets, but am hoping to contain it within a reasonable space.

Berkeman, the functions of the electromagnet and the DAC would overlap with some other functions of the experiment as a whole :grumpy:. Turns out the sample needs to be moved magnetically.
Last edited:
  • #7
You can get Nd magnets off the shelf; peruse the inventory
of these places and you'll generally find several available
varieties: [Broken]

e.g. [Broken]

click here for color photo Mighty Micro Magnet tell a friend about this item so they can check it out
Absurd strength in this rare earth magnet proves once again that size doesn't matter. Half the size of your pinky nail, the 3/16" x 3/8" x .05" thick magnets can lift 7 oz. of steel and are polarized on their ends not on their flat sides. It's .5 grams and strongest on its ends. Great for use with reed switches. Very, very bad to have around credit cards.
on sale 32436 MICROMAGNET $1.50 / PKG(4)
(was $2.00) add to cart

click here for color photo Itty Bitty Shiny Bar Magnet tell a friend about this item so they can check it out
This neodymium bar is a slight 1/4" x 19/32" x 1/8" thick, but strong for its size. Little as it is, according to our strength and fitness lab, it can lift 3 lbs. Nice polished silver color, too.
on sale 37408 NEO BAR MAGNET $2.95 / PKG(3)
(was $3.75) [Broken]

Cylindrical Lift tell a friend about this item so they can check it out
A good alternative for when you need to fit a magnet into a little hole, like in a door closure. These neodymium magnets are 13/32" dia x 19/32" long, with approx 4 lbs of lift. Cute and useful. Sold in packages of (2) or for larger savings try the package of (20).
37403P2 4-LB CYLINDER MAGNET $3.95/PKG(2) add to cart
37403P20 4-LB CYLINDER MAGNET $19.50/PKG(20) add to cart
Cylindrical Lift
click here for color photo [Broken]

click here for color photo Powerful Flat Disc tell a friend about this item so they can check it out
Proportionally, this little neodymium magnet is way, way, way stronger than you. Stronger than you and your favorite football team, to tell the truth, since this little guy is only 1-1/2" dia x 1/4" thick, and he has a 25-lb pull. With a 3/16" hole in his center!
37404 NEO DISK MAGNET $5.95 / EACH add to cart
Last edited by a moderator:
  • #8
Hey thanks a bunch XEZ!

This is great stuff! Lets see what I can do with these. Should get my string's attention finally. I'll check back with some more questions in a couple of days.

Great job, Thanks again!
  • #9
Since your thread has a small diameter, you'll want to
use a magnet (or pole piece) that is sufficiently small
or geometrically tapered / pointed in height so that the
field is very highly spatially non-uniform in the vicinity of
the thread and pole piece; that way the attraction force
will be strongest due to the change in field energy vs.
change of thread position.

I think using some kind of mechanical actuator and
an appropriate high strength Nd magnet moving in
proximity to the sample may be most effective in
field strength variance while being more compact in size
than an electromagnet.

However on the topic of electromagnets, it looks like there
are various high saturation field ferrites
(e.g. various Manganese Zinc) that have
saturation field densities of between 0.4 and 0.58 Tesla
or so.

There is another material called 'MPP' Molypermalloy
Mo-Ni-Fe which saturates around 0.75T.

There's another material, Kool Mu which saturates
around 1.05T.

There's another material called 'High Flux' that saturates
around 1.5T.

The 'High Flux' may be your best bet due to high
saturation flux density as well as permeability,
remanence, et. al. though MPP wouldn't be a bad
choice, and KoolMu probably better than most ferrites.

You'd have to get either an 'E' core or cylinder or
half of a pot core, or saw/break a toroid for an
electromagnet. Wind it so that at your nominal coil
current the Amp-Turns gives enough Oersteds of (H)
in the core that H * mu will give you a nearly
saturated flux density and that's the best that you're
going to do with that size/material of core.

You'll still need to supply a fairly substantial amount of
energy quickly to rapidly switch the magnetization
of the cores, though, all the way from saturation on
one side of their curve into saturation on the other side.
The benefit of using the right material is that you get
high field strengths at at least a somewhat beneficial
permeability factor by having the core there versus
having just air. Core Inductor/PowderCore_Guide.php [Broken][1].pdf PDF files/6.pdf

et. al.

(Molypermalloy Mo-Ni-Fe),
Last edited by a moderator:
  • #10

That one got some neurons seriously fired up! :D Going to take me a while to peruse the catalogs. I'll do that and get this working soon. Will inform you on how it turns out.

Thanks a bunch again!!!

Would give you some stars if i could.

See you in a bit.
  • #11
You're quite welcome; good luck!

Just keep in mind that when you work with strong
rare earth magnets, you must assume that you're going
to have zero physical control of them anywhere within a
couple of inches of a ferromagnetic substance or another

They're ceramic, so they will chip / shatter easily if
they're allowed to have things clunk against them.

You're likely going to be unable to peel them off direct
contact with a ferromagnet or other magnet without
massive difficulty.

So usually it's best to keep them well away from such
objects, and when you're intentionally going to bring
them close do so only if they're physically distanced by
at least 0.1" or so of intervening strong plastic/wooden

To make some kind of apparatus where they are intended
to be close to another magnet or ferromagnetic material,
you'd devise some kind of mounting mechanism for the
individual magnet where it's clamped in place firmly but
gently (because they're shatter like glass is the pressure
or abrasion is too high in any spot). Then once the
individual magnet is clamped to a holder, you can use some
mechanism like a screw / vise / whatever to adjust
the distance from the magnet holder to get the close
proximity you need while being able to control / adjust
the spacing and prevent things from launching themselves

You'll get the best field concentration with a design
where a north pole is very close to a south pole and in
between those is your field area. Join the opposite
'unused' pole ends via a conduit of iron if possible
so there'll be a high permeability flux path in a fairly
complete circle with the primary air-gap being your
sample area.

It may work ok without a nearly closed high permeability
magnetic path, and with using only one magnet instead
of two facing pole pieces, but if you need more field
intensity in the area, using the gap between
paired pole pieces, is always an option to consider.
  • #12
Hey XEZ,

I was thinking - are the cores that you mentioned - MPP/High Flux/Sendust available only as circular cores? Because that would mean that the field lines would be completely within the core itself right? Would it stick to an iron surface if the wire is wound in a toroidal fashion? Are these available as straight rods, instead of just circular cores? I think you understand what I'm getting to - I'd need a field that comes out of the solenoid fairly significantly. And how hard is it to break a large MPP core, so that I get a horseshoe type core which I can then use?

The price for an MPP 0.380 inch O.D. powdered core is $0.80. Do they really have such low hysteresis effects? I imagined that for high permeability (low amp-turns), low hysteresis material, the price would be really high, unless maybe this is no big deal. Can I use this ( ) for my switching function without the residual effects?

At the moment, I believe this magnet would serve my purposes since it is small enough that for a low amp - say 1 A, I'd get a pretty decent magnetic field : 0.05 T maybe for around 23 turns for 20G. The sample and the magnet would be separated by some medium, maybe wood or acrylic.

Am I missing something? Somehow sounds too good to be true - maybe the fact that the field stays mostly within for a toroidal winding. If i got a thicker core and wind the wires around the circumference, what would the direction of the field be given a hollow center?

Is there any way that I can measure how 'far' the magnetic field would actually be strong, i mean - theoretically. For how much distance is this Flux density maintained?

Also, there are a couple of terms that I don't understand from the catalogs:

1. In the specs for these cores, they mention Turns and Single Layer turns - whats the difference? and which should I consider?
2. Whats "u.w.f" ?
3. Whats the Mean magnetic Path Length?
4. Is there some page listing the hysteresis loss coefficient for these material so that I can calculate the Core loss, along with the other losses?

Also, passing this amount of current for a winding this thick - the resistance being low, wouldn't that heat the winding up? Say, for a longer pulse.

Thanks again,

This experiment just got more interesting.
Last edited:

Related Threads on Solenoid Electromagnet with strong core, low residual magnetism

  • Last Post
  • Last Post
  • Last Post
  • Last Post